公开(公告)号：US12177282B2

公开(公告)日：2024-12-24

申请号：US17955444

申请日：2022-09-28

Applicant: Samsung Electronics Co., Ltd.

Inventor： Luckraj Shrawan Kumar ,  Gopinath Chennakeswaran ,  Renuka Varry ,  Divyanshu Chuchra ,  Umang Bhatia ,  Utkarsh Mohan ,  Aalok Kumar Gupta

IPC: H04L65/756 ,  H04L65/75

Abstract: An example method includes for enhancing video streaming includes inputting a client-side environment parameter to a client-side neural network (NN) implemented in a client device and then training the client-side NN based on the inputted client-side environment parameter. Thereafter, a client-side encoding bit rate is determined based on a result of the training of the client-side NN and then transmitting the determined client-side encoding bit rate as an input to a server-side neural network (NN) implemented in a server device. The method further includes training the server-side NN based on the inputted client-side encoding bit rate and a server-side environment parameter. Then, a server-side encoding bit rate is determined based on a result of the training of the server-side NN and thereafter, an encoding bit rate of an encoder for the video stream is adjusted based on the determination of the server-side encoding bit rate, and the server-side environment parameter.

公开(公告)号：US20240385742A1

公开(公告)日：2024-11-21

申请号：US18659965

申请日：2024-05-09

Applicant: Samsung Electronics Co., Ltd.

Inventor： Nazrinbanu Nurmohammad NAGORI ,  Divyanshu Chuchra ,  Jiyoun Lee

IPC: G06F3/0488 ,  H04N7/01

Abstract: Embodiments herein provide a method for user interaction management in a user-interactive application in an electronic device. The method includes: generating and displaying artificial frame in-between a plurality of real frames of the user-interactive application and receiving a user input on the artificial frame displayed in-between the plurality of real frames of the user-interactive application; detecting an object in the artificial frame on which the user input is received; determining a correlation between the user input, coordinates of object in real frame of the plurality of real frames, and coordinates of the object in the artificial frame based on the detected object in the artificial frame; and displaying an interaction of the object in the artificial frame based on the determined correlation.

公开(公告)号：US10997795B2

公开(公告)日：2021-05-04

申请号：US16409243

申请日：2019-05-10

Applicant: Samsung Electronics Co., Ltd.

Inventor： Raghavan Velappan ,  Suresh Kumar KrishnanKutty Vettukuzhyparambhil ,  Pavan Kumar Dusi ,  Raghavendra Holla ,  Amit Yadav ,  Nachiketa Das ,  Divyanshu Chuchra

IPC: G06T19/20 ,  G06T7/90 ,  G06T7/60 ,  G06T15/08 ,  G06T9/00

Abstract: An apparatus and method are provided for compressing a three-dimensional (3D) object image represented by point cloud data. The method includes positioning the 3D object image into a plurality of equi-sized cubes for compression; determining 3D local coordinates in each of the plurality of equi-sized cubes and a cube index for each point of the 3D object image positioned in the plurality of equi-sized cubes; generating two-dimensional (2D) image data based on the 3D local coordinates and the cube indexes; and storing the 2D image data in a memory. The 2D image data includes at least one of 2D geometry data, 2D meta data, or 2D color data.